A peptide binding motif for I-Eg7, the MHC class II molecule that protects E alpha-transgenic nonobese diabetic mice from autoimmune diabetes.

The nonobese diabetic (NOD) mouse, a model of spontaneous insulin-dependent diabetes mellitus (IDDM), fails to express surface MHC class II I-Eg7 molecules due to a deletion in the E alpha gene promoter. E alpha-transgenic NOD mice express the E alpha E beta g7 dimer and fail to develop either insulitis or IDDM. A number of hypotheses have been proposed to explain the mechanisms of protection, most of which require peptide binding to I-Eg7. To define the requirements for peptide binding to I-Eg7, we first identified an I-Eg7-restricted T cell epitope corresponding to the sequence 4-13 of Mycobacterium tuberculosis 65-kDa heat shock protein (hsp). Single amino acid substitutions at individual positions revealed a motif for peptide binding to I-Eg7 characterized by two primary anchors at relative position (p) 1 and 4, and two secondary anchors at p6 and p9. This motif is present in eight of nine hsp peptides that bind to I-Eg7 with high affinity. The I-Eg7 binding motif displays a unique p4 anchor compared with the other known I-E motifs, and major differences are found between I-Eg7 and I-Ag7 binding motifs. Analysis of peptide binding to I-Eg7 and I-Ag7 molecules as well as proliferative responses of draining lymph node cells from hsp-primed NOD and E alpha-transgenic NOD mice to overlapping hsp peptides revealed that the two MHC molecules bind different peptides. Of 80 hsp peptides tested, none bind with high affinity to both MHC molecules, arguing against some of the mechanisms hypothesized to explain protection from IDDM in E alpha-transgenic NOD mice.     

Complexity among constituents of the HLA-B*1501 peptide motif

 Analysis of peptides derived from HLA class I molecules indicates that thousands of unique peptides are bound by a single molecular type, and sequence examination of the pooled constituents yields a motif which collectively defines the peptides bound by a given class I molecule. Motifs resulting from pooled sequencing are then used to infer whether particular viral and tumor protein fragments might serve as class I-presented peptide therapeutics. Still undetermined from a pooled motif is the breadth or range of peptides in the population which are brought together to form the pooled motif, and it is therefore not yet known how representative of the population a pooled motif is. By employing hollow fiber bioreactors for large-scale production of HLA class I molecules, sufficient peptides are produced to investigate individual subsets of peptides comprising a motif. Edman sequencing and mass spectrometric analysis of peptides eluted from HLA-B^*1501 reveal that many peptide sequences fail to align with either the N- or C-terminal anchors predicted for the B^*1501 peptide motif through whole pool sequencing. These analyses further reveal auxiliary anchors not previously detected and peptides significantly larger and smaller than the predicted nonamer, ranging from 6 to 12 amino acids in length. These results demonstrate that constituents of the B^*1501 peptide pool vary markedly in comparison with one another and therefore in comparison with previously established B^*1501 motifs, and such complexity indicates that many of the peptide ligands presented to CTL cannot be predicted using class I consensus motifs as search criteria. Received: 7 October 1997 / Revised: 10 December 1997     

Allelic variation in key peptide-binding pockets discriminates between closely related diabetes-protective and diabetes-susceptible HLA-DQB1*06 alleles

HLA-DQA1*0102-DQB1*0602 is associated with protection against type 1 diabetes (T1D). A similar allele, HLA-DQA1*0102-DQB1*0604, contributes to T1D susceptibility in certain populations but differs only at seven amino acids from HLA-DQA1*0102-DQB1*0602. Five of these polymorphisms are found within the peptide-binding groove, suggesting that differences in peptide binding contribute to the mechanism of their association with T1D. In this study, we determine the peptide-binding motif for HLA-DQA1*0102-DQB1*0604 allelic protein (DQ0604) in comparison to the established HLA-DQA1*0102-DQB1*0602 (DQ0602) motif using binding assays with model peptides from T1D autoantigens and homology modeling using the coordinates of the DQ0602-hypocretin 1-13 crystal structure. The peptide binding preferences were deduced with a peptide from insulin that bound both with a 2- to 3-fold difference in avidity using the same amino acids in the peptide as anchors. Peptide binding differences directly influenced by the polymorphisms in or nearby pockets 1, 6, and 9 were observed. In pocket 1, DQ0604 was better able to accommodate aromatic residues due to the beta86 and beta87 polymorphisms. A negatively charged amino acid was preferred by DQ0604 in pocket 6 due to the positively charged beta30His. In pocket 9, DQ0604 preferred aromatic amino acids due to the beta9 and beta30 polymorphisms and had low tolerance of acidic residues. beta57Val in DQ0604 functions differently than beta57Ala, in that it pushes alpha76Arg outside of the pocket, preventing the formation of a salt bridge with an acidic amino acid in the peptide. This study furthers our understanding of the structure-function relationships of MHC class II polymorphisms.     

Functional classification of class II human leukocyte antigen (HLA) molecules reveals seven different supertypes and a surprising degree of repertoire sharing across supertypes

Previous studies have attempted to define human leukocyte antigen (HLA) class II supertypes, analogous to the case for class I, on the basis of shared peptide-binding motifs or structure. In the present study, we determined the binding capacity of a large panel of non-redundant peptides for a set of 27 common HLA DR, DQ, and DP molecules. The measured binding data were then used to define class II supertypes on the basis of shared binding repertoires. Seven different supertypes (main DR, DR4, DRB3, main DQ, DQ7, main DP, and DP2) were defined. The molecules associated with the respective supertypes fell largely along lines defined by MHC locus and reflect, in broad terms, commonalities in reported peptide-binding motifs. Repertoire overlaps between molecules within the same class II supertype were found to be similar in magnitude to what has been observed for HLA class I supertypes. Surprisingly, however, the degree to which repertoires between molecules in the different class II supertypes also overlapped was found to be five to tenfold higher than repertoire overlaps noted between molecules in different class I supertypes. These results highlight a high degree of repertoire overlap amongst all HLA class II molecules, perhaps reflecting binding in multiple registers, and more pronounced dependence on backbone interactions rather than peptide anchor residues. This fundamental difference between HLA class I and class II would not have been predicted on the basis of analysis of either binding motifs or the sequence/predicted structures of the HLA molecules.     

Redundancy in antigen-presenting function of the HLA-DR and -DQ molecules in the multiple sclerosis-associated HLA-DR2 haplotype

The three HLA class II alleles of the DR2 haplotype, DRB1*1501, DRB5*0101, and DQB1*0602, are in strong linkage disequilibrium and confer most of the genetic risk to multiple sclerosis. Functional redundancy in Ag presentation by these class II molecules would allow recognition by a single TCR of identical peptides with the different restriction elements, facilitating T cell activation and providing one explanation how a disease-associated HLA haplotype could be linked to a CD4+ T cell-mediated autoimmune disease. Using combinatorial peptide libraries and B cell lines expressing single HLA-DR/DQ molecules, we show that two of five in vivo-expanded and likely disease-relevant, cross-reactive cerebrospinal fluid-infiltrating T cell clones use multiple disease-associated HLA class II molecules as restriction elements. One of these T cell clones recognizes >30 identical foreign and human peptides using all DR and DQ molecules of the multiple sclerosis-associated DR2 haplotype. A T cell signaling machinery tuned for efficient responses to weak ligands together with structural features of the TCR-HLA/peptide complex result in this promiscuous HLA class II restriction.     

The HLA molecules DQA1*0501/B1*0201 and DQA1*0301/B1*0302 share an extensive overlap in peptide binding specificity

Assays to measure the binding capacity of peptides for HLA-DQA1*0501/B*0201 (DQ2.3) and DQA1*0301/B*0302 (DQ3.2) were developed using solubilized MHC molecules purified from EBV-transformed cell lines. These quantitative assays, based on the principle of the inhibition of binding of a high-affinity radiolabeled ligand, were validated by examining the binding capacity of known DQ-restricted epitopes or ligands. The availability of these assays allowed an investigation of patterns of cross-reactivity between different DQ molecules and with various common DR molecules. DQ2.3 and DQ3.2 were found to have significantly overlapping peptide binding repertoires. Specifically, of 13 peptides that bound either DQ2.3 or DQ3.2, nine (69.2%) bound both. The molecular basis of this high degree of cross-reactivity was further investigated with panels of single substitution analogs of the thyroid peroxidase 632-645Y epitope. It was found that DQ2.3 and DQ3.2 bind the same ligands by using similar anchor residues but different registers. These data suggest that in analogy to what was previously described for HLA-DR molecules, HLA-DQ supertypes characterized by largely overlapping binding repertoires can be defined. In light of the known linkage of both HLA-DQ2.3 and -DQ3.2 with insulin-dependent diabetes mellitus and celiac disease, these results might have important implications for understanding HLA class II autoimmune disease associations.     

Improved prediction of MHC class I and class II epitopes using a novel Gibbs sampling approach

MOTIVATION: Prediction of which peptides will bind a specific major histocompatibility complex (MHC) constitutes an important step in identifying potential T-cell epitopes suitable as vaccine candidates. MHC class II binding peptides have a broad length distribution complicating such predictions. Thus, identifying the correct alignment is a crucial part of identifying the core of an MHC class II binding motif. In this context, we wish to describe a novel Gibbs motif sampler method ideally suited for recognizing such weak sequence motifs. The method is based on the Gibbs sampling method, and it incorporates novel features optimized for the task of recognizing the binding motif of MHC classes I and II. The method locates the binding motif in a set of sequences and characterizes the motif in terms of a weight-matrix. Subsequently, the weight-matrix can be applied to identifying effectively potential MHC binding peptides and to guiding the process of rational vaccine design. RESULTS: We apply the motif sampler method to the complex problem of MHC class II binding. The input to the method is amino acid peptide sequences extracted from the public databases of SYFPEITHI and MHCPEP and known to bind to the MHC class II complex HLA-DR4(B1*0401). Prior identification of information-rich (anchor) positions in the binding motif is shown to improve the predictive performance of the Gibbs sampler. Similarly, a consensus solution obtained from an ensemble average over suboptimal solutions is shown to outperform the use of a single optimal solution. In a large-scale benchmark calculation, the performance is quantified using relative operating characteristics curve (ROC) plots and we make a detailed comparison of the performance with that of both the TEPITOPE method and a weight-matrix derived using the conventional alignment algorithm of ClustalW. The calculation demonstrates that the predictive performance of the Gibbs sampler is higher than that of ClustalW and in most cases also higher than that of the TEPITOPE method.     

Analysis of MHC class II antigen processing by quantitation of peptides that constitute nested sets

Peptides associated with class II MHC molecules are of variable length because in contrast to peptides associated with class I MHC molecules, their amino and C termini are not constrained by the structure of the peptide interaction with the binding site. The proteolytic processing events that generate these peptides are still not well understood. To address this question, peptides extracted from HLA-DR*0401 were analyzed using two types of mass spectrometry instrumentation. This enabled identification of >700 candidate peptides in a single analysis and provided relative abundance information on 142 peptides contained in 11 nested sets of 3-36 members each. Peptides of 12 residues or less occurred only at low abundance, despite the fact that they were predicted to fully occupy the HLA-DR*0401 molecule in a single register. Conversely, the relative abundance of longer species suggested that proteolytic events occurring after MHC binding determine the final structure of most class II-associated peptides. Our data suggest that C-terminal residues of these peptides reflect the action of peptidases that cleave at preferred amino acids, while amino termini appear to be determined more by proximity to the class II MHC binding site. Thus, the analysis of abundance information for class II-associated peptides comprising nested sets has offered new insights into proteolytic processing of MHC class II-associated peptides.     

Androgen receptor ligand-binding domain interaction and nuclear receptor specificity of FXXLF and LXXLL motifs as determined by L/F swapping

The androgen receptor (AR) ligand-binding domain (LBD) binds FXXLF motifs, present in the AR N-terminal domain and AR-specific cofactors, and some LXXLL motifs of nuclear receptor coactivators. We demonstrated that in the context of the AR FXXLF motif many different amino acid residues at positions +2 and +3 are compatible with strong AR LBD interaction, although a preference for E at +2 and K or R at +3 was found. Pairwise systematic analysis of F/L swaps at +1 and +5 in FXXLF and LXXLL motifs showed: 1) F to L substitutions in natural FXXLF motifs abolished AR LBD interaction; 2) binding of interacting LXXLL motifs was unchanged or increased upon L to F substitutions; 3) certain noninteracting LXXLL motifs became strongly AR-interacting FXXLF motifs; whereas 4) other nonbinders remained unaffected by L to F substitutions. All FXXLF motifs, but not the corresponding LXXLL motifs, displayed a strong preference for AR LBD. Progesterone receptor LBD interacted with some FXXLF motifs, albeit always less efficiently than corresponding LXXLL motifs. AR LBD interaction of most FXXLF and LXXLL peptides depended on classical charge clamp residue K720, whereas E897 was less important. Other charged residues lining the AR coactivator-binding groove, K717 and R726, modulated optimal peptide binding. Interestingly, these four charged residues affected binding of individual peptides independent of an F or L at +1 and +5 in swap experiments. In conclusion, F residues determine strong and selective peptide interactions with AR. Sequences flanking the core motif determine the specific mode of FXXLF and LXXLL interactions.     

Large scale mass spectrometric profiling of peptides eluted from HLA molecules reveals N-terminal-extended peptide motifs

The majority of >2000 HLA class I molecules can be clustered according to overlapping peptide binding specificities or motifs recognized by CD8(+) T cells. HLA class I motifs are classified based on the specificity of residues located in the P2 and the C-terminal positions of the peptide. However, it has been suggested that other positions might be relevant for peptide binding to HLA class I molecules and therefore be used for further characterization of HLA class I motifs. In this study we performed large-scale sequencing of endogenous peptides eluted from K562 cells (HLA class I null) made to express a single HLA molecule from HLA-B*3501, -B*3502, -B*3503, -B*3504, -B*3506, or -B*3508. Using sequence data from >1,000 peptides, we characterized novel peptide motifs that include dominant anchor residues extending to all positions in the peptide. The length distribution of HLA-B35-bound peptides included peptides of up to 15 residues. Remarkably, we determined that some peptides longer than 11 residues represented N-terminal-extended peptides containing an appropriate HLA-B35 peptide motif. These results provide evidence for the occurrence of endogenous N-terminal-extended peptide-HLA class I configurations. In addition, these results expand the knowledge about the identity of anchor positions in HLA class I-associated peptides that can be used for characterization of HLA class I motifs.     

An HLA-DRB1-derived peptide associated with protection against rheumatoid arthritis is naturally processed by human APCs

Predisposition to rheumatoid arthritis (RA) is thought to be associated with HLA-DR1, -DR4, and -DR10. However, many epidemiological observations are better explained by a model in which the DQ alleles that are linked to these DR alleles, i.e., DQ5, DQ7, and DQ8, predispose to RA, while certain DR alleles have a dominant protective effect. All protective DRB1 alleles, e.g., *0402, *1301, and *1302, encode a unique motif, (70)DERAA(74). The protection may be explained by the presentation of DRB1-derived peptides by DQ to immunoregulatory T cells, because it was demonstrated in various autoimmune disease models that T cell responses to certain self-Ags can be involved in disease suppression. The aim of this study was to analyze whether peptides carrying the DERAA motif are naturally processed by human APC and presented in the context of the RA-predisposing DQ. Using a synthetic peptide carrying the DRB1*0402-derived sequence (65)KDILEDERAAVDTYC(79), we generated DERAA peptide-specific DQ-restricted T cell clones (TCC) from a DQ8 homozygous individual carrying DERAA-negative DR4 alleles. By analyzing the proliferation of these TCC, we demonstrated natural processing and presentation of the DERAA sequence by the APC of all the individuals (n = 12) carrying a DERAA-positive DRB1 allele and either DQ8 or the DQ8-related DQ7. Using a panel of truncated synthetic peptides, we identified the sequence (67)(I)LEDERAAVD(TY)(78) as the minimal determinant for binding to DQ8 and for recognition by the TCC. These findings support a model in which self-MHC-derived peptide can modulate predisposition to autoimmune disease in humans.     

Differential induction by immune interferon of the gene products of the HLA-D region on the melanoma cell line MeWo and its metastatic variant MeM 50-10

This study has shown for the first time an association between the metastatic properties of two autologous melanoma cell lines and their susceptibility to induction of HLA class II Ag by IFN-gamma. After in vitro incubation with IFN-gamma the melanoma cell line MeWo did not acquire reactivity with anti-HLA class II antibodies, whereas its metastatic variant MeM 50-10 did. The differential susceptibility to induction of HLA class II Ag on the two cell lines cannot be accounted for by either differences in the number and affinity of IFN-gamma receptors or in the sensitivity to IFN-gamma, but most likely reflects an intrinsic property of each cell line. Serologic and immunochemical investigations with anti HLA-DR, DQ, and DP mAb have indicated that only HLA-DR Ag are induced by IFN-gamma on MeM 50-10 cells. Northern blot analysis with HLA-DR beta, DQ beta, and DP beta probes suggest that different mechanisms underlie the differential susceptibility to induction by IFN-gamma of the gene products of the HLA-D region. The regulatory mechanism(s) that control the expression of HLA class II Ag appear to be different from those controlling the expression of the melanoma-associated Ag tested, inasmuch as the modulation of the latter by IFN-gamma did not differ on the two melanoma cell lines.     

Naturally processed HLA class II peptides reveal highly conserved immunogenic flanking region sequence preferences that reflect antigen processing rather than peptide-MHC interactions

MHC class II heterodimers bind peptides 12-20 aa in length. The peptide flanking residues (PFRs) of these ligands extend from a central binding core consisting of nine amino acids. Increasing evidence suggests that the PFRs can alter the immunogenicity of T cell epitopes. We have previously noted that eluted peptide pool sequence data derived from an MHC class II Ag reflect patterns of enrichment not only in the core binding region but also in the PFRS: We sought to distinguish whether these enrichments reflect cellular processes or direct MHC-peptide interactions. Using the multiple sclerosis-associated allele HLA-DR2, pool sequence data from naturally processed ligands were compared with the patterns of enrichment obtained by binding semicombinatorial peptide libraries to empty HLA-DR2 molecules. Naturally processed ligands revealed patterns of enrichment reflecting both the binding motif of HLA-DR2 (position (P)1, aliphatic; P4, bulky hydrophobic; and P6, polar) as well as the nonbound flanking regions, including acidic residues at the N terminus and basic residues at the C terminus. These PFR enrichments were independent of MHC-peptide interactions. Further studies revealed similar patterns in nine other HLA alleles, with the C-terminal basic residues being as highly conserved as the previously described N-terminal prolines of MHC class II ligands. There is evidence that addition of C-terminal basic PFRs to known peptide epitopes is able to enhance both processing as well as T cell activation. Recognition of these allele-transcending patterns in the PFRs may prove useful in epitope identification and vaccine design.     

T cell immunity to type II collagen in the biobreeding rat: the identification and characterization of RT1u-restricted T cell epitopes on alpha 1(II)

Susceptibility to experimental collagen-induced arthritis in rodents is dependent on MHC class II elements to bind peptides from the type II collagen (CII) molecule. Although a substantial body of data has been reported in mice defining these peptide Ags, little has been reported in rats. In this study, we investigate the locations and sequences of CII peptides, which are bound by RT1(u) molecules, expressed by diabetic-resistant, arthritis-susceptible Biobreeding rats, and, in turn, stimulate CII-specific T cells. By using overlapping and substituted peptide homologues of CII, we have identified and characterized an immunodominant and five subdominant epitopes on CII, which stimulate RT1(u)-restricted T cell proliferation. The immunodominant epitope, CII (186-192), contains a QGPRG core sequence, which was found in a subdominant epitope CII (906-916). Similar sequences containing single conservative substitutions were identified in three other epitopes. One, CII (263-272), contained a conservatively substituted R-->K substitution, whereas CII (880-889) and CII (906-916) contained nonconservative substitutions, i.e., P-->D and R-->M, respectively. Homologue peptides containing these sequences stimulated T cell proliferative responses, although less intensely than peptides containing CII (186-192). Substituting QGR residues in the QGPRG core with alanine, isoleucine, or proline reduced proliferation, as did substituting flanking E and G residues at the N terminus and E at the C terminus. Collectively, these data indicate that RT1(u)-restricted immunodominant and several subdominant epitopes on CII often share a QGPRG-like motif, with conservative substitutions present at either P or R positions. This motif is similar to one recognized by collagen-induced arthritis-susceptible HLA-DR1- and HLA-DR4-transgenic mice.     

Editing of the HLA-DR-peptide repertoire by HLA-DM.

Antigenic peptide loading of classical major histocompatibility complex (MHC) class II molecules requires the exchange of the endogenous invariant chain fragment CLIP (class II associated Ii peptide) for peptides derived from antigenic proteins. This process is facilitated by the non-classical MHC class II molecule HLA-DM (DM) which catalyzes the removal of CLIP. Up to now it has been unclear whether DM releases self-peptides other than CLIP and thereby modifies the peptide repertoire presented to T cells. Here we report that DM can release a variety of peptides from HLA-DR molecules. DR molecules isolated from lymphoblastoid cell lines were found to carry a sizeable fraction of self-peptides that are sensitive to the action of DM. The structural basis for this DM sensitivity was elucidated by high-performance size exclusion chromatography and a novel mass spectrometry binding assay. The results demonstrate that the overall kinetic stability of a peptide bound to DR determines its sensitivity to removal by DM. We show that DM removes preferentially those peptides that contain at least one suboptimal side chain at one of their anchor positions or those that are shorter than 11 residues. These findings provide a rationale for the previously described ligand motifs and the minimal length requirements of naturally processed DR-associated self-peptides. Thus, in endosomal compartments, where peptide loading takes place, DM can function as a versatile peptide editor that selects for high-stability MHC class II-peptide complexes by kinetic proofreading before these complexes are presented to T cells.     

A Viral,Transporter Associated with Antigen Processing (TAP)-independent,High Affinity Ligand with Alternative Interactions Endogenously Presented by the Nonclassical Human Leukocyte Antigen E Class I Molecule

The transporter associated with antigen processing (TAP) enables the flow of viral peptides generated in the cytosol by the proteasome and other proteases to the endoplasmic reticulum, where they complex with nascent human leukocyte antigen (HLA) class I. Later, these peptide-HLA class I complexes can be recognized by CD8⁺ lymphocytes. Cancerous cells and infected cells in which TAP is blocked, as well as individuals with unusable TAP complexes, are able to present peptides on HLA class I by generating them through TAP-independent processing pathways. Here, we identify a physiologically processed HLA-E ligand derived from the D8L protein in TAP-deficient vaccinia virus-infected cells. This natural high affinity HLA-E class I ligand uses alternative interactions to the anchor motifs previously described to be presented on nonclassical HLA class I molecules. This octameric peptide was also presented on HLA-Cw1 with similar binding affinity on both classical and nonclassical class I molecules. In addition, this viral peptide inhibits HLA-E-mediated cytolysis by natural killer cells. Comparison between the amino acid sequences of the presenting HLA-E and HLA-Cw1 alleles revealed a shared structural motif in both HLA class molecules, which could be related to their observed similar cross-reactivity affinities. This motif consists of several residues located on the floor of the peptide-binding site. These data expand the role of HLA-E as an antigen-presenting molecule.     

Identification of immunodominant epitopes derived from the respiratory syncytial virus fusion protein that are recognized by human CD4 T cells

Memory CD4 T-cell responses against respiratory syncytial virus (RSV) were evaluated in peripheral blood mononuclear cells of healthy blood donors with gamma interferon enzyme-linked immunospot (Elispot) assays. RSV-specific responses were detected in every donor at levels varying between 0.05 and 0.3% of CD4 T cells. For all donors tested, a considerable component of the CD4 T-cell response was directed against the fusion (F) protein of RSV. We characterized a set of 31 immunodominant antigenic peptides targeted by CD4 T cells in the context of the most prevalent HLA class II molecules within the Caucasian population. Most antigenic peptides were HLA-DR restricted, whereas two dominant DQ peptides were also identified. The antigenic peptides identified were located across the entire sequence of the F protein. Several peptides were presented by more than one major histocompatibility complex class II molecule. Furthermore, most donors recognized several F peptides. Detailed knowledge about immunodominant antigenic peptides will facilitate the ability to monitor CD4 T-cell responses in patients and the measurement of correlates of protection in vaccinated subjects.     

Quantitative predictions of peptide binding to any HLA-DR molecule of known sequence: NetMHCIIpan

CD4 positive T helper cells control many aspects of specific immunity. These cells are specific for peptides derived from protein antigens and presented by molecules of the extremely polymorphic major histocompatibility complex (MHC) class II system. The identification of peptides that bind to MHC class II molecules is therefore of pivotal importance for rational discovery of immune epitopes. HLA-DR is a prominent example of a human MHC class II. Here, we present a method, NetMHCIIpan, that allows for pan-specific predictions of peptide binding to any HLA-DR molecule of known sequence. The method is derived from a large compilation of quantitative HLA-DR binding events covering 14 of the more than 500 known HLA-DR alleles. Taking both peptide and HLA sequence information into account, the method can generalize and predict peptide binding also for HLA-DR molecules where experimental data is absent. Validation of the method includes identification of endogenously derived HLA class II ligands, cross-validation, leave-one-molecule-out, and binding motif identification for hitherto uncharacterized HLA-DR molecules. The validation shows that the method can successfully predict binding for HLA-DR molecules-even in the absence of specific data for the particular molecule in question. Moreover, when compared to TEPITOPE, currently the only other publicly available prediction method aiming at providing broad HLA-DR allelic coverage, NetMHCIIpan performs equivalently for alleles included in the training of TEPITOPE while outperforming TEPITOPE on novel alleles. We propose that the method can be used to identify those hitherto uncharacterized alleles, which should be addressed experimentally in future updates of the method to cover the polymorphism of HLA-DR most efficiently. We thus conclude that the presented method meets the challenge of keeping up with the MHC polymorphism discovery rate and that it can be used to sample the MHC "space," enabling a highly efficient iterative process for improving MHC class II binding predictions.